S.Figure 3

Genetic correlations (A) across trait levels [factor scores] and (B) trait variability [quantile integrted rank scores]

SNP based genetic correlations estimated using LDSC from summary statistics following GWA (A) and vGWA (B) of childhood behavioural traits in the MoBa sample; Point-size=Z-statistic; Colour is strength of genetic correlation (Red=positive; Blue=negative)

A Factor scores

B Quantile integrated ranks scores

S.Figure 3. Genetic correlation matrices for mGWA (A) and vGWA (B) shows the degree to which genetic effects are shared across the phenotypes. The colour intensity represents the direction and strength of the genetic correlation (red = positive correlation, white = no correlation, blue = negative correlation) and the size of each point represents the z-statistic (rg / se). For mGWA (left), the correlation structure is more dense than for vGWA (right). This is likely because of differences in statistical power for detecting mean and variance effects - as a much smaller proportion of vGWA summary statistics had statistically significant h2 estimates.